Vaccination has played a key role in the control of viral diseases during the past 30 years. Vaccination is based on a simple principle of immunity: once exposed to an infectious agent, an animal mounts an immune defense that provides lifelong protection against disease caused by the same agent. The goal of vaccination is to induce the animal to mount the defense prior to infection. Conventionally, this has been accomplished through the use of live attenuated or whole inactivated forms of the virus as immunogens. The success of these approaches depends on the presentation of native antigen which elicits the complete range of immune responses obtained in natural infection.
Despite their considerable success, conventional vaccine methodologies are subject to a number of potential limitations. Insufficiently inactivated vaccines may cause the disease they are designed to prevent. Attenuated strains can mutate to become more virulent or non-immunogenic. Viruses that can establish latency, such as the herpesviruses, are of particular concern as it is not known whether there are any long-term negative consequences of latent infection by attenuated strains. Finally, there are no efficient means of growing many types of viruses.
Recent advances in recombinant DNA technology offer the potential for developing vaccines based on the use of defined antigens as immunogens, rather than the intact infectious agent. These include peptide vaccines, consisting of chemically synthesized, immunoreactive epitopes; subunit vaccines, produced by expression of viral proteins in recombinant heterologous cells; and the use of live viral vectors for the presentation of one or more defined antigens.
Both peptide and subunit vaccines are subject to a number of potential limitations. A major problem is the difficulty of ensuring that the conformation of the engineered proteins mimics that of the antigens in their natural environment. Suitable adjuvants and, in the case of peptides, carrier proteins, must be used to boost the immune response. In addition these vaccines elicit primarily humoral responses, and thus may fail to evoke effective immunity. Subunit vaccines are often ineffective for diseases in which whole inactivated virus can be demonstrated to provide protection. For example, canine parvovirus subunits fail to elicit virus-neutralizing antibodies in rabbits (Smith and Halling, Gene, 29:263-269 (1984)), although protective inactivated vaccines are available.
As an alternative to recombinant-produced subunit vaccines comprising a purified polypeptide, it may be possible to develop non-infectious, subunit-like vaccines that consist of viral capsid proteins assembled into virus-like structures. Such non-replicating, virus-like particles would have many of the immunologic advantages of inactivated vaccines combined with the safety features of subunit vaccines. Several researchers have reported the development of eukaryotic systems for the expression of foreign viral capsid proteins, and the self assembly of these proteins into virus-like particles. For example, co-expression of canine parvovirus (CPV) capsid proteins VP1 and VP2 in murine cells transformed with a bovine papilloma virus/CPV recombinant plasmid resulted in the formation of self-assembling particles that resembled, biochemically and immunologically, authentic CPV virions (Mazzara, et al., Modern Approaches to Vaccine, Cold Spring Harbor Laboratory, N.Y., R. M. Chanock and R. A. Lerner, eds. pp. 419-424 (1986); Mazzara, et al., PCT Application No. WO88/02026, published Mar. 24, 1988). When used to vaccinate susceptible dogs, these empty capsids elicited immune responses capable of protecting against CPV challenge. In another example, it has been shown that the expression of HIV or SIV gag precursor polypeptide in insect cells using the baculovirus expression system results in the formation of immature, retroviral-like particles that are secreted into the culture medium of infected cells (Gheysen, et al., Cell, 59:103-112 (1989); Delchambre, et al., EMBO J., 8:2653-2660 (1989)). In mammalian cells, HIV-like particles that contained core polypeptides as well as reverse transcriptase were produced after transient expression of the HIV gag-pol genes using an SV40 late replacement vector (Smith, et al., J. Virol., 64:2653-2659 (1990)).
Recombinant vaccinia viruses that express at least the HIV gag gene have also been shown to give rise to the production of retroviral-like particles upon infection of appropriate host cells (Karacostas, et al., Proc. Natl. Acad. Sci. USA, 86:8964-8967 (1989); Shiota and Shibuta, Virology, 175:139-148 (1990)). The coexpression in recombinant vaccinia-infected cells of gag polypeptides with the HIV envelope glycoproteins resulted in the formation of HIV-like particles that comprised an enveloped core structure containing, embedded in the envelope, the HIV envelope glycoproteins. The coexpression of gag and env genes in infected cells could be achieved by co-infecting the cells with two different recombinant vaccinia viruses, one expressing env and one expressing gag-pol (Haffar, et al., J. Virol., 64:2653-2659 (1990)), or by infecting the cells with a single recombinant that expressed both env and gag-pol (Mazzara, et al., U.S. patent application Ser. Nos. 07/360,027 and 07/540,109).
The ability to produce particles containing viral envelope glycoproteins has important implications for vaccine development. Viral envelope glycoproteins, which are located in the outer lipid membrane of enveloped viruses (such as herpesviruses, retroviruses, togaviruses, rhabdoviruses, paramyxoviruses, orthomyxoviruses and coronaviruses) are often the major immunogenic determinants of the virus. In the case of HIV, for example, the envelope glycoprotein gp120 contains the key epitopes that elicit virus-neutralizing antibody responses (Arthur, L. A., et al., Proc. Natl. Acad. Sci. USA, 84:8583-8587 (1987)). Similarly, the herpes simplex virus glycoprotein gB and the rabies glycoprotein both elicit virus-neutralizing antibody responses and, in addition, have been shown to protect against challenge with the cognate pathogens in the absence of other viral proteins (Paoletti, et al., Proc. Natl. Acad. Sci. USA, 81:193-197 (1984); Wiktor, et al., Proc. Natl. Acad. Sci. USA, 81:7194-7198 (1984)).
Unfortunately, there are many viruses for which heterologous expression of self-assembling viral capsids may not prove feasible. Formation of herpesviruses capsids, for example, would require the expression of more genes than can be practically accommodated in available expression vectors. The mechanism of particle assembly for a number of other viruses, such as the helical RNA viruses, makes self assembly of virus-like particles from a heterologous expression system problematic. Nonetheless, it would be useful to be able to produce non-infectious, self-assembling virus-like particles containing membrane glycoproteins from any enveloped virus.
Envelope glycoproteins from viruses of different families can be incorporated at low frequency into heterologous virus particles by the biological phenomenon known as pseudotyping or phenotypic mixing. In co-infection experiments, the genome of one virus species can be demonstrated to be physically associated with glycoproteins from the other species. In a review of the literature on this phenomenon, Zavada, (J. Gen. Virol., 63:15-24 (1982)) cites examples of pseudotyping between, for example, retroviruses and togaviruses, rhabdoviruses, paramyxoviruses or herpesviruses. For pseudotyping to occur, the two viruses must have compatible life cycles, i.e., neither must interfere with the replication of the other. Recently, Zhu, et al., (J. Acquired Immune Deficiency Syndromes, 3:215-219 (1990)) described phenotypic mixing between HIV and vesicular stomatitis virus or herpes simplex virus.